• Created 6 comprehensive machine learning online courses, which enrolled 50k+ students and received 10k+ positive reviews.• Created Feature-engine, an open-source Python package for feature engineering and selection, with 150k+ monthly downloads, 1.3k+ stars on Github, and a 40+ contributor community.• Authored the "Python Feature Engineering Cookbook" for Packt Publishing.• Authored the book "Feature Selection in Machine Learning with Python," available on… Show more • Created 6 comprehensive machine learning online courses, which enrolled 50k+ students and received 10k+ positive reviews.• Created Feature-engine, an open-source Python package for feature engineering and selection, with 150k+ monthly downloads, 1.3k+ stars on Github, and a 40+ contributor community.• Authored the "Python Feature Engineering Cookbook" for Packt Publishing.• Authored the book "Feature Selection in Machine Learning with Python," available on Leanpub.• Founded Train in Data, a school with intermediate and advanced courses on data science and machine learning.• Winner of the Data Leaders Awards 2018 for Data Science and Analytics• Selected one of LinkedIn Voices 2019 for data science and analytics. Show less

Researched, developed and put into production machine learning models for Motor Claims and Fraud Prevention. Lead in the implementation of Machine Learning across the business. Mentored data scientists and supported in the recruitment of members for the data science team.

• Built machine-learning models for credit risk, fraud and collections.• Built and curated SQL reporting code following compliance.• Used Python (scikit-learn, pandas, scipy, numpy, matplotlib, keras) and SQL (PostgreSQL and Microsoft SQL).• Statistical hypothesis testing for different PoC implementations.

• Generated insight and designed solutions through open source software and algorithms following briefs from clients, including identification of trends and patterns and building of predictive models (ARIMA models for forecasting and machine learning algorithms: regression, trees, SVMs).• Delivered presentations to and collaborated with senior business stakeholders.• Employed a variety of models (linear and logistic regression, ARIMA models, SVMs, trees) and tools (Python, R and… Show more • Generated insight and designed solutions through open source software and algorithms following briefs from clients, including identification of trends and patterns and building of predictive models (ARIMA models for forecasting and machine learning algorithms: regression, trees, SVMs).• Delivered presentations to and collaborated with senior business stakeholders.• Employed a variety of models (linear and logistic regression, ARIMA models, SVMs, trees) and tools (Python, R and SQL). Show less

• Secured funds for research (£800,000+) by writing 2 research programmes.• Led research studies that were published in the world’s 3rd most cited multidisciplinary scientific journal (Galli et al., Nat Commun, 2014).• Determined the impact of drugs on neuronal connections using parametric (e.g. anova, t-test) and non-parametric (e.g., Kruskal-Wallis, Mann-Whitney U test) statistical tests.• Using k-means clustering, identified 2 populations of neurons in a brain area… Show more • Secured funds for research (£800,000+) by writing 2 research programmes.• Led research studies that were published in the world’s 3rd most cited multidisciplinary scientific journal (Galli et al., Nat Commun, 2014).• Determined the impact of drugs on neuronal connections using parametric (e.g. anova, t-test) and non-parametric (e.g., Kruskal-Wallis, Mann-Whitney U test) statistical tests.• Using k-means clustering, identified 2 populations of neurons in a brain area that respond differently to treatment with a specific molecule.• Presented data at universities (UK, Finland, Turkey and Argentina) and in international conferences in Europe and the US.• Contributed to media dissemination of research findings (TheNewScientist, News Medical and Parkinson’s UK charity website). Links below.List of publications:https://www.pubfacts.com/author/Soledad+Galli Show less

• Contributed to the development of a novel algorithm to determine and quantify colocalisation in fluorescence images (link below).• Supervised 3 MSc students and mentored junior staff.• Wrote 4 articles for publication in international scientific journals, one of them recommended in the Faculty of 1000, a renowned biomedical site dedicated to help scientists discover and discuss research.